User blog:Homuhomu123/Shelling Accuracy Test Results
Foreword This post is used for saving experiment data from other Kancolle testers, on shelling accuracy during combat. For a higher level of objectivity, I will refrain from concluding or formulating from anything below. In order to help readers understand the exp't, I attached some information beside the original data according to my understanding of the experiments. Therefore you're always welcome to evaluate my work & translation. Since the following tests are all about probability, data from smaller sample sizes would be surrounded by parenthesis, indicating less reliability. For more information on the significance of larger sample sizes, you may visit Calculating Confidence Intervals for Underlying Probabilities by Winzongod. Since some info below were pasted from websites with no explicit citation, please contact me if you know any of the sources. Also, if you consider any unauthorized using of exp't data, you'll have my apology and associated content shall be deleted immediately. If you noticed an error on my translation, ambiguity in the language, any important info were missing, or just need help with anything, please leave me a comment. I hope the following will be helpful. I - Tentative Shelling Accuracy Formula (2014) Reports / Exp't name: *「艦これ」砲撃命中率について・その3「損傷の影響」 *「艦これ」砲撃命中率について・索敵の影響 *「艦これ」砲撃命中率について・単縦陣まとめ *「艦これ」砲撃命中率について・装備の影響 Author: ??? (Arranged & Posted on 半製造BSｱﾚｺﾚ雑記帳) Date: June ~ October, 2014 Explanation: ''' A beta formula regarding day battle shelling accuracy (single attack) had been derived, based on earlier experiments data from players. According to the exp't findings, shelling accuracy equals to the ''Accuracy Term'' minus ''Evasion Term''. The 2 terms are affected by attacking ship's level, equipment bonus on accuracy, and evasion stat of the target. ''.. Accuracy Term = 1 + sqrt( (Attacker's Lv - 1) *5 ) / 100 + Equipment_Bonus / 100'' ''.. Evasion Term = Evasion / 75 .... for Evasion ≦ 37.5：'' '' = Evasion / (Evasion + 37.5) .... for Evasion ＞37.5：'' :: >> Shelling Accuracy = Accuray Term - Evasion Term :: >> Dodge Rate = 100% - Shelling Accuracy '''Accuracy Term: As the attacking ship's level grows, ceteris paribus, its accuracy term increases from 100% to 122% (with diminishing return). The offense fleet uses Line Ahead. Derived from test ships with level 1, level 3 and level 90 only, thus not necessarily accurate. Evasion Term: For ships with low evasion, dodge rate positively & linearly relates to its evasion stat. When evasion stat passes 37.5, diminishing return expected. Attacking fleet uses Line Ahead. Not necessarily accurate. Limitation: The above formula, however, had its own constraints due to the following reasons: * Luck modifier '''was not considered. * '''Line Ahead formation only. * Has to be in full morale (not sparkled) * Overweight penalty & fit gun bonus not included. * Day battle, single attacks only * Ship type (carrier??) The 4 sub-sections below serve as experimental basis for the tentative formula. The last sub-section is a revised formula by the author in recognition of the luck modifier (the existence of which is to be further verified). (a) Health State & Shelling Accuracy source: [http://bs-arekore.at.webry.info/201407/article_5.html '「艦これ」砲撃命中率について・その3「損傷の影響」] A fleet with level 3 ships were sent to map 2-1-1. Hits from BOTH sides were recorded along with their health. Data compared & combined with previous experiments. : Since the abyssal ships had very similar evasion stats (12. 14. 15. 16), thus the shelling accuracy on which will not be distinguished. : Despite smaller samples from the attacking ships (chuuha, taiha), it seems that: *Enemy's health state will not affect your ship's accuracy. *Your ship's accuracy cannot be lower significantly when moderately / severly damaged. (Assume the attacking ship won't gain accuracy when considerably damaged) : '' '' (b) Success in Detection Phase & Shelling Accuracy ''source: 「艦これ」砲撃命中率について・索敵の影響 Placing a high level BB / SS as flagship, fully loaded with seaplane scouts to ensure the success in detection. The other ships are low level ships for test purpose. (either Lv.1 or 3). *Line Ahead *Map 3-2-1 or 2-1-1 *"Which ship attacks which" was specifically recorded. *Attacks from the flagship, of course, is not included : Data were compared with previous experiments, where the fleet falied to detect the enemy (due to low FLoS). : Despite smaller samples from successful detection, this factor seems to be irrelevant with shelling accuracy. . © Verification of the Shelling Accuracy Formula (2014) 'source: ''' 「艦これ」砲撃命中率について・単縦陣まとめ The formula is now being verified based on previous experiment data. Those tests were all performed under the following condition: *Maps: 3-2-1, 2-1-1, 3-3-1 *Morale 47~49. *Ship levels: Lv.1, 3 or ~90, *No accuracy bonus from equipment. *Using formation Line Ahead. *"Which ship attacks which" was specifically recorded. The following conditions were NOT '''firmly controlled: *Ship type *Engagement form *Health state *Success in detection phase (d) Effects from Equipment Bonus on Accuracy ''source: 「艦これ」砲撃命中率について・装備の影響 A fleet of Lv.91~92 ships were sent to map 3-3-1, Each had equipment(s) giving +8 bonus on accuracy. Note that BB(V) was not included in this test thus "fit gun modifer" was not a concern. *Using formation Line Ahead *Neither fatigued nor sparkled. *DD, CL, CA, CAV. *Success in detection guaranteed. Despite of high ship level (~90), the +8 bonus from equipment did provide ~8% bonus on tested accuracy. (e) Shelling Accuracy Formula Revised (2015) source: 「艦これ」砲撃命中率について・最新の推定式 Date: Before April, 2015 Explanation: As more data were collected, the old accuracy formula is to be revised for better estimation. The new formula (2015, April) recognizes the luck modifier (existence of which is to be verified). The results, however, are not very different from the previous one. It was tested that the shelling accuracy of a ship, ceteris paribus, increases by 20% (addition) from level 1 to 99. Married ship, however, seems to have additional accuracy which cannot be described by the following formula. Luck modifer does exist, might be slightly higher than "luck / 10". This formula is still under constant review thus subject to change. .. Accuracy Term = 1 + ( sqrt( Attacker's Lv ) - 1 ) / 45 + Equipment_Bonus / 100 + Luck / 1000 .. Evasion Term = Evasion / 75 .... for Evasion ≦ 37.5： '' = Evasion / (Evasion + 37.5) .... for Evasion ＞37.5：'' :: >> Shelling Accuracy = Accuray Term - Evasion Term :: >> Dodge Rate = 100% - Shelling Accuracy II - Double Line vs Line Ahead Reports / Exp't name: '''「艦これ」砲撃命中率について・複縦陣まとめ '''Author: ??? (Arranged & Posted on 半製造BSｱﾚｺﾚ雑記帳) Date: January, 2015 Explanation: ''' : Same test condition as section I, except choosing Double Line (DL) instead of Line Ahead (LA) formation. : This experiment serves as a comparison to the previous tests, from which the accuracy bonus from DL could be identified & measured. : All accuracy below pertain to hit rates on the abyssal fleet. : : '''PART (A): In the first part (map 3-2-1, friendly ship level = Lv1), we witnessed an increase of ~20% on shelling accuracy. *''Y-Axis: accuracy'' *''X-Axis: evasion stat (abyssal)'' *The blue dots represent experimental accuracy obtained by using Doule Line. *The red dots represents those by using Line Ahead. *The vertical streches are the confidence interval for each data point. (95%) *Green dots are the differences. : PART (B): : In the second part'' (map 2-1-1, friendly ship level Lv.3), we witnessed an '''accuracy cap' at around 97%. : To illustrate, the hit rates for Line Ahead was 81.8 ~ 86.2%, which only increased to ~97% when using Double Line. : III - Preliminary Test on the Overweight Penalty (Kongou) Report / Exp't name: ??? Author / Blog: ??? '''(Repasted on PPT BBS / 看板KanColle ) '''Date: Summer, 2014 Explanation: The abuse of 46cm triple main gun on Kongou class had prevailed since the beginning of Kancolle. They were, however, unable to use Armor Piercing shells. According to a certain update in the summer of 2014, all fast battleships gained access to AP shell. As compensation, the "overweight penalty" was introduced. Kancolle twitter announced that "Overloading main gun considered too heavy for the BB would induce penalty on accuracy, while using fit gun on them will gain accuracy bonus." By which extent those penalty / bonus were, however, was still a myth. As a result, the following test regarding the overweight mechanics on FBB has been conducted (source unknown). Test map is believed to be 1-1-1. Date is assumed to be late summer since the report was pasted as early as September, 2014. Despite that, test ship conditions are all given in the chart. Also, please be aware of the accuracy cap (~97%). Simplified Test Results Original Data IV - The "Fit Gun + Overweight" Mechanics on BB (Red Morale Test) Report / Exp't name: VARIOUS '' (命中 / 1-1-1 / 赤疲労検証)'' Author / Blog: ''VARIOUS '' *''Posted by''' 36butm520 @ Shitaraba BBS [1] [2].'' *''Arranged & summarized on 艦これ 検証＆攻略 Wiki*'' '''Date: Since Fall, 2014 Explanation: : All the experiment data were conducted by unknown players. Test results posted on 2 threads of Shitaraba BBS [1] [2].Some of them were also arranged on wikiwiki.jp , along with other experiment data. : Test map 1-1-1. Ships were mostly Lv 90+. In order to avoid hitting the accuracy cap (~97%), test ships were kept in red morale to deliberately lower their accuracy. : For easier viewing on the equipment set up, the following icons will be used: : ' _ ....' Empty slot : ▲ …. Fit Gun. (eg., 46cm on Yamato) : X ''.... ''Overweight Gun'' (eg. 41cm on Kongou)'' : R ''.... ''Radar : >> Since the true probablity are around 50%, tests with sample sizes lower than 300 are less reliable, thus will NOT be shown in the table below. Results from sample > 400 are considered more convincing. You may check the original posts to see the complete data. Simplified Test Results Original Data *''Please visit 1-1 赤疲労検証(623) / (624)' to check them.'' *''Some results could be also found on '艦これ 検証＆攻略 Wiki*.'' V - Effects of Morale on Accuracy (& Evasion) Report / Exp't name: ''VARIOUS'' Author / Blog: VARIOUS ('Arranged & summarized on 艦これ 検証＆攻略 Wiki*)'' '''Date: Sep.~ Nov., 2014 Explanation: : 53 was known as the step value for your ship to gain "double" evasion & accuracy. According to some reports, KC players came to realize there might be another step value between 30~39, for the accuracy & evasion penalty to come to effect. : Meanwhile, people were also trying to figure out how the "fit gun" modifiers apply to shelling accuracy. It had been suspected that those bonus / overweight penalty, were some sort of fixed values independent of other factors. Simplified Test Results *'Morale & Accuracy' *'Morale & Evasion' VI - Luck stat & Shelling Accuracy Report / Exp't name: ''雪風改と五月雨改で1-1-1命中検証'' Author / Blog: ??? ('Posted on Shitaraba BBS, Arranged on 艦これ 検証＆攻略 Wiki*)'' 'Date: ' October 25, 2014 '''Explanation: : Luck stat had been recognized as a main factor in the occurence of night battle cut-in. That is being said, it had no known effects in combat. It was believed by some people that luck stat relates to accuracy. This test was conducted to verify that hypothesis. : Test map 1-1-1. Morale 33 ~ 49. Test ships are Samidare, Miyuki & Yukikaze (x2). Since they leveled up during the experiment, the average will be displayed in the chart. Simplified Test Results Note: %Crit = critical hits / attack chance. VII - Accuracy Bonus from Improved Equipment Report / Exp't name: ''VARIOUS'' ('装備改修の命中検証 + 命中/1-1-1/無疲労) 'Author / Blog: ''VARIOUS ('Posted on Shitaraba BBS, ''''Arranged on 艦これ 検証＆攻略 Wiki*) Date: Since November, 2014 Explanation: : According to player reports on wikiwiki, improved equipment (from Akashi's Improvement Arsenal) gives additional firepower / torpedo. It is also believed that these improvement provides hidden bonus on accuracy. : Test map 1-1-1. In order to avoid hitting the accuracy cap (~97%), some tests used the red morale testing method. Accuracy from such tests will be shown in RED. : The other experiment used normal morale testing method (cond 33~49), given ship levels were much lower (~30). The average ship level during the whole process will be considered. For ships with extremely low levels (1+), their levels shal be distinguished specifically. Simplified Test Results *Upgrade: (★max) Note: %Crit = critical hits / attack chance. *Upgrade: (★1) *Upgrade: (★4) *Map: 2-1-1 VIII - Night Battle DA Accuracy Bonus by using 20.3cm (CA.CAV) Report / Exp't name: ''20.3主砲 VS 15.5副砲'' Author / Blog: ??? Date: Fall, 2014 Explanation: ''' Kancolle twitter posted on Sept.18, 2014: ''>07▼Battle related stats improved for CA and its variation! '' ''Greater accuracy and evasion. Bonus from 20.3cm (and its variation). '' This study was conducted to verify that, especially night battle DA accuracy bonus from using 20.3cm on CA(V). A total of 100 battles, map 5-3-C (night battle). The author divided test ships into 2 fleets using different calibur guns: * '''Fleet No.1: ''<~ 20.3cm x2'' ** Haguro Lv51, Kako Lv46, Ashigara Lv44, Atago Lv45, Kumano Lv50, Kitakami Lv98 * Fleet No.2:'' <~ 15.5cm(sec) x2'' ** Takao Lv50, Furutaka Lv46, Myoukou Lv43, Nachi Lv46, Suzuya Lv50, Ooi Lv98 Equipment on the remaining slots: * For CA: Type 3 shell + #21 Radar * For CAV: Prototype Seiran (+1 acc) + #21 Radar * For CLT: Type A Ko-hyoteki Test ship condition: *beyond moderate damage *morale = 49 Simplified Test Results Original Data IX - The "Fit Gun + Overweight" Mechanics on CL Report / Exp't name: ''VARIOUS'' Author / Blog: VARIOUS ('Posted on Shitaraba BBS)'' '''Date: Since May, 2015 Explanation: 「艦これ」開発/運営 @KanColle_STAFF 09▼軽巡洋艦主砲【フィット補正】追加更新 軽巡洋艦の主砲運用において、よりフィットした主砲装備の運用による命中率補正等が強化されます。※特に複数装備する場合、小型軽量の主砲装備が有効です。 #艦これ After a certain update in May 2015, the "Fit Gun" bonus now applies on CL. The twitter says "Fit guns on Light Cruisers now provide accuracy bonus, especially when multiple equipment are used on the ship." Although there was no hint for the "Overweight Penalty" on the official twitter, players started testing if stacking 20.3cm (and variations) would lower the accuracy of a CL. Similar to the previous BB overweight testing. Map 1-1-1 or 3-2-1* (only I-class & Ro-class DDs were considered in calculation). Ships were mostly Lv 90+. In order to avoid hitting the accuracy cap (~97%), test ships were kept in red morale to deliberately lower their accuracy. Category:Blog posts